topic Re: AZEO vs 100% w/balances? in Understanding FICO® Scoring

AZEO vs 100% w/balances?

Anonymous — Wed, 27 Jun 2018 04:44:25 GMT

Looking for data points from those with clean files that have tested going from AZEO to 100% of cards with balances reported. I'm curious how many points were lost in going from this "best" to "worst" number of revolvers with balances reported. If anyone took down data points along the way (different percentage of cards with balances reported) it would be cool to hear those.

For those that do chime in, it would be crucial to know how many total revolvers we're talking. Profiles with 3+ revolvers would be best in terms of testing. I know TT posts a table from time to time where he went from 0 to I believe all 6 of his revolvers with balances reported, but with his scores maxed at 850 and buffer being part of the equation I'm not sure we could really tell how many points were lost in going to 100% of cards with balances reported.

Anyway, I was at AZEO a few weeks back and have since allowed 2 more of my 8 revolvers to report balances. I'm going to let the other 5 report balances as well over the next few weeks, so around this time next month I'll see the impact of 100% of revolvers with balances. From there I'll step down card by card reporting $0 again, just to double check my numbers from the progression. This will be my first time on a clean file testing this. On my dirty file it didn't matter if I was at AZEO or 100% of cards with balances reported, as my scores remained constant.

Re: AZEO vs 100% w/balances?

Anonymous — Wed, 27 Jun 2018 05:54:37 GMT

Thanks BBS!

If anyone does wish to post results, I strongly encourage him to post:

(a) The number of open revolving accounts (at zero and with a positive balance)

(b) The number of open installment accounts (presumably all with a positive balance)

I have written about this several times before, but any serious look at this needs to include this information as well. FICO's negative reason statements typically refer to the "number of accounts" showing a balance -- not ratios and not revolving only. I have always suggested that ratios might be one of the ways FICO looks at this, but there's no reason to think it is the only thing that matters. In the last year the "ratio" idea has taken on a life of its own and through repetition has crowded out all other possibilities, even at the expense of the plain wording of the reason codes.

As I have mentioned before I will sometimes get the "too many accounts with a balance" negative reason statement when I have < 20% of my cards showing a balance but an integer number of 4-5 accounts with a balance (counting loans).

Many thanks to BBS for giving us his actual number of cards. I hope he'll also remind us how many open loans he has as well.

Re: AZEO vs 100% w/balances?

Subexistence — Wed, 27 Jun 2018 13:40:56 GMT

@Anonymous wrote:
Thanks BBS!

If anyone does wish to post results, I strongly encourage him to post:
(a) The number of open revolving accounts (at zero and with a positive balance)
(b) The number of open installment accounts (presumably all with a positive balance)

I have written about this several times before, but any serious look at this needs to include this information as well. FICO's negative reason statements typically refer to the "number of accounts" showing a balance -- not ratios and not revolving only. I have always suggested that ratios might be one of the ways FICO looks at this, but there's no reason to think it is the only thing that matters. In the last year the "ratio" idea has taken on a life of its own and through repetition has crowded out all other possibilities, even at the expense of the plain wording of the reason codes.

As I have mentioned before I will sometimes get the "too many accounts with a balance" negative reason statement when I have < 20% of my cards showing a balance but an integer number of 4-5 accounts with a balance (counting loans).

Many thanks to BBS for giving us his actual number of cards. I hope he'll also remind us how many open loans he has as well.

Does AZEO really mean AZEtwo for people with SSL? Like for me, ideally I get my BBR and SSL to report each month and they are both accounts with balances.

Re: AZEO vs 100% w/balances?

HeavenOhio — Wed, 27 Jun 2018 15:09:05 GMT

AZEO refers to revolving accounts. FICO does give a "too many accounts with balances" reason, which may refer to all accounts. But as you know, all cards reporting zero will ding you. The most practical scenario for most people to achieve their best scores would be to let one card report a small positive balance and let the loans and reason statements fall where they fall.

Re: AZEO vs 100% w/balances?

Anonymous — Wed, 27 Jun 2018 15:11:08 GMT

Hello SubEx! AZEO is a strategy, and like all strategies it is based on what you can control. If you have two car loans and a mortgage (say) you cannot keep them open while reducing some of them to a zero balance. The three loans will have three positive balances and there is no strategy that will change that.

If you have several credit cards, however, you can keep all at zero except one (and still keep them open). Since card balances are controllable, AZEO focuses only on card balances (including LOCs, etc.).

So to answer your (explicit) question, AZEO only refers to cards.

To answer your implicit question, however, it is quite possible that the back-end FICO algorithm counts open loans as "accounts showing a balance." (The language of the negative reason statement.) The reason to think that the algorithm might do that is that they ARE accounts with a balance. And because we have seen some actual evidence along that line.

We know, for example, that at least one person (can't remember his name) has experienced absolutely no FICO 8 effect on going from exactly one card to 100% of all their cards showing a balance. When I probed a bit deeper with that fellow, however, I found that he had something like nine open loans and exactly three credit cards. Thus he went from 9 out of 12 open accounts showing a balance to 12 out of 12. And then we have my own experience where I receive that negative reason statement with a very low percentage of cards showing a balance, but an integer number of (say) five.

Bear in mind too that the different scoring models could each work differently in this regard. FICO 8 is a different model from FICO 9 which is different from FICO 04 (TU and EQ mortgage) which is different from FICO 98 (EX mortgage).

The ideal test subjects in my mind would be two classes of people:

(a) Lots of loans and exactly three cards.

(b) Many cards (at least 8 but 12+ would be even better) and exactly 1 open loan.

B is common but A is also commoner than you'd think at first. Class A most often consists of people coming out of college with a bunch of student loans and possibly a car loan.

Re: AZEO vs 100% w/balances?

Anonymous — Wed, 27 Jun 2018 15:29:35 GMT

I fall into "b" above... 8 revolvers and just a single open installment loan (mortgage) to refresh the memory of all.

I believe the last time I saw the "too many accounts with balances" negative reason statement was back in Feb/March when I also had an open auto loan. At AZEO on my revolvers, that means that only 3 accounts had balances against 10 total open accounts. I don't believe I've seen that statement since then, when dropping to 2 of now 9 total accounts with balances. Interesting, though, that I have had 2 revolvers report a balance (AZE2) so I have been at 3 "accounts" with balances reported as well as 4 total (AZE3 + mortgage).

Unfortunately, CCT doesn't provide me any negative reason statements. I only get them monthly from a couple of my lenders like Synchrony (TU FICO 8) and my mortgage (TU FICO 4). Perhaps as my number of accounts with balances increases during my test I'll see that statement pop back up. Since I only get those scores/statements once every 30 days from a few of my lenders, I can't do a strategic pull at a point in time that I'd like the way I can with multiple CCT trials. When those monthly accounts do update, I just see the generation date and cross reference that to my CCT pull dates to know the exact data the algorithm was looking at when those reason statements were generated.

I would like to try and time it so that all 8 of my revolvers have balances at the time I get my TU FICO 4 mortgage score pull/update from my mortgage account. It would be cool to be able to actually quantify the variance between the score drop from my TU FICO 8 score at AZEO vs 100% of cards with balances and my TU FICO 4 score with the same change. We commonly hear on this forum that, "mortgage scores are more sensitive to number of cards with balances" but rarely does anyone quantify the difference. It could be 5 points or it could be 50 points and I would have no idea. I'd like to know and plan to have an answer about a month from now.

Good point from CGID in the original reply to this thread that number of non-revolving accounts present is an important factor for anyone contributing here to include.

Re: AZEO vs 100% w/balances?

Anonymous — Wed, 27 Jun 2018 16:05:48 GMT

Great plan, BBS. We might want to drop into the student loan forum (and the general forum) and explain that we are looking for people wishing to test this. And that we are looking to find people of class A or B. (The SL forum because we are looking for B.)

My own tentative guess (in lieu of hard data, involving repeated tests by a multiple people) is that FICO models use both metrics: percentage of accounts and integer number. I am also agnostic on whether installment accounts are considered -- I am just saying that experiments have to be designed to test that possibility.

SouthJ would be an interesting guy for this, since he has 21+ cards. We could compare what happens when he goes from AZEO to just under 33% of cards or accounts reporting. If he sees a penalty but you do not (same ratio for you) it would suggest that integer number also matters.

Comparing people in class A vs. class B would be a different test: an easy way to see whether FICO 8 considers loans in the "number of accounts with a balance" metrics.

Re: AZEO vs 100% w/balances?

Subexistence — Wed, 27 Jun 2018 17:53:22 GMT

Are you going to ask people who did SSL to participate as type B?

Re: AZEO vs 100% w/balances?

Anonymous — Wed, 27 Jun 2018 18:12:00 GMT

@Subexistence wrote:
Are you going to ask people who did SSL to participate as type B?

Sure. SSLs (Share Secured Loans) and SLs (Student Loans) and mortgages and auto loans (etc.) are all examples of loans. Either FICO counts all loans in its "number of accounts reporting a balance" or it counts no loans. I suppose one could imagine a model that counted auto loans but ignored personal loans (say) but that would be a really odd choice to make.

The reason I mentioned the student loan forum as a good place to look for Group B is that recent college grads with who are frequenting a room to ask about student loans are the most likely people to have eight different open loans. In theory a person could have four auto loans and three personal loans and a mortgage, but that's going to be very rare.

Re: AZEO vs 100% w/balances?

Anonymous — Wed, 27 Jun 2018 18:22:33 GMT

The ideal member of group B is someone with many loans (8+) and exactly three cards (no other revolving accounts). That way he is guaranteed to be at > 80% of his open accounts showing a balance (9 accounts) even while having cards at AZEO.

If FICO does count loans, such a person will likely see no score penalty going from AZEO to 100% of his cards, since he started with almost all of his accounts showing a balance anyway (9 out of 11).

If FICO does not count loans, then we'd likely see a score jump -- assuming FICO 8 cares about number of accounts at all.

Re: AZEO vs 100% w/balances?

Anonymous — Wed, 27 Jun 2018 20:29:13 GMT

CGID, do you think it's possible though that both matter? Perhaps the person with a bunch of loans and only 3 revolvers may already be at a high number (and percentage) of accounts with non-zero balances, but could just revolvers be looked at from a percentage perspective independent of loans? I'm not sure there's a concrete answer to this. If so, perhaps someone with 33% of revolvers that goes from AZEO to 3/3 (100%) could experience a slight ding still?

Re: AZEO vs 100% w/balances?

Anonymous — Wed, 27 Jun 2018 22:54:09 GMT

@Anonymous wrote:
CGID, do you think it's possible though that both matter? Perhaps the person with a bunch of loans and only 3 revolvers may already be at a high number (and percentage) of accounts with non-zero balances, but could just revolvers be looked at from a percentage perspective independent of loans? I'm not sure there's a concrete answer to this. If so, perhaps someone with 33% of revolvers that goes from AZEO to 3/3 (100%) could experience a slight ding still?

@I think it is very possible that both things matter. That's why I wrote:

"I have always suggested that ratios might be one of the ways FICO looks at this, but there's no reason to think it is the only thing that matters."

My only concern has been the way that people talk about this as if the sole metric would the "percent of cards." As long as that is the thunderously omnipresent meme, there's no way to explore whether or not it is entirely true.

Re: AZEO vs 100% w/balances?

Subexistence — Wed, 27 Jun 2018 23:38:29 GMT

What about the people with SSL just being a part of the group that has many revolvers and 1 loan.

Re: AZEO vs 100% w/balances?

Anonymous — Wed, 27 Jun 2018 23:41:22 GMT

@Subexistence wrote:
What about the people with SSL just being a part of the group that has many revolvers and 1 loan.

I don't understand your question.

Re: AZEO vs 100% w/balances?

Anonymous — Wed, 27 Jun 2018 23:48:10 GMT

@Anonymous
My only concern has been the way that people talk about this as if the sole metric would the "percent of cards." As long as that is the thunderously omnipresent meme, there's no way to explore whether or not it is entirely true.

I find it odd that people look at it that way, since the exact lingo provided by FICO reason statements is "number" and not "percentage."

Re: AZEO vs 100% w/balances?

jamie123 — Thu, 28 Jun 2018 01:07:28 GMT

I have tested this many times in the past and will test it again this fall when I set up an AZEO situation.

I can't remember the points lost for each additional reporting card because if I remember correctly it varied a bit and was around 5 points a card. I do however remember that there was a 15 point difference when crossing the 50% of cards reporting a balance threshold. I was testing this when I had 8 CCs and 1 SSL. If I let 4 cards report a small balance I lost 15 points. If I only let 3 cards report a balance I would gain the 15 points back.

I will retest this in the fall. I will have 10 CCs, 1 mortgage and 1 auto loan open this time around.

Re: AZEO vs 100% w/balances?

Subexistence — Thu, 28 Jun 2018 01:20:45 GMT

@Anonymous wrote:
@Subexistence wrote:
What about the people with SSL just being a part of the group that has many revolvers and 1 loan.
I don't understand your question.

The ideal test subjects in my mind would be two classes of people:

(a) Lots of loans and exactly three cards.

(b) Many cards (at least 8 but 12+ would be even better) and exactly 1 open loan.

I meant that some people with only 1 installment loan, the SSL loan, should be in category B described by many cards and exactly 1 open loan.

Re: AZEO vs 100% w/balances?

Anonymous — Thu, 28 Jun 2018 01:21:23 GMT

Jamie, from your post above it suggests that you saw point losses with each additional revolver reporting a balance above AZEO. Is that really what you saw on your profile? If so, that's a bit different than what is commonly believed to be the case, where certain thresholds exist.

Re: AZEO vs 100% w/balances?

Thomas_Thumb — Thu, 28 Jun 2018 01:31:31 GMT

I saw something like what Jamie reported on my EQ BCE Fico 8 (below graph posted previously in other threads).

Re: AZEO vs 100% w/balances?

Anonymous — Thu, 28 Jun 2018 01:36:06 GMT

I don't doubt that with the BCE model, which is known to be more sensitive to cards with balances. With Classic 8 though, this is something that I haven't heard of yet.